One of the most thorough and elegant demonstrations of the stability of IQ was conducted by a group of scientists in New Zealand in the late 1960s, although that is not what they set out to study. Obstetric medicine and care had improved dramatically in the previous decades. Many more babies were surviving childbirth, particularly babies who experienced problems just prior to and during delivery. Doctors began to worry, however, that these children were prone to greater problems as they grew up. Perhaps by decreasing infant mortality, new medical practices had led to increased childhood morbidity. The theory was that the problems that might have caused difficulties during birth were now creating problems later down the road when these babies were a bit older. In an effort to track the well-being of this new population of babies, a group of doctors and psychologists in New Zealand decided to try to follow a large cohort of babies all born in one place at one time. The researchers followed the fate of 1,037 babies, all born at the Queen Mary Hospital in Dunedin between April 1, 1972, and March 31, 1973. Almost all of the babies chosen for the study have been visited, observed, and tested every few years right up to the present. Many of those babies now have babies of their own who are being observed and interviewed. One of the extraordinary features of this endeavor is that so many of the children remained in the study.
As anyone who has ever done longitudinal research will tell you, the hardest part is keeping your subjects in the study. Most studies of this kind suffer from attrition. Say a scientist recruits one hundred babies for her study. By the time the babies are ten years old, fifty of the families might have moved away, twenty more might refuse to be part of follow-up assessments, and before you know it, the researcher has only thirty subjects to work with and, as a result, can say almost nothing about broad trends. So it is particularly impressive that when the Dunedin kids were twenty-one years old, almost all of them were still around and happy to answer the researchers' questions, take the required tests, and participate in interviews. Ninety-three percent participated in a daylong assessment. Only nineteen refused to participate, and another nine couldn't be tracked down.
In the Dunedin study, doctors were able to learn a great deal about the health patterns of the children as they grew up. They began to see which kinds of illnesses were the sequelae of improved pre- and perinatal care. But as it turned out, they also learned quite a lot about the children's intelligence.
They first assessed the children's IQ when they were young, as part of a complete battery of baseline measures. Then, to be consistent, they measured the children's IQ every three years, through adolescence and early adulthood. And here's the simple but startling fact: when it came to performance on an IQ score, the overwhelming majority of the children (930 out of 1,037) stayed pretty much the same. The researchers first thought that they had found straightforward evidence that IQ does not change. However, the beauty and difficulty of large data sets is that you always get some noise.